Got a Suspicious Message? Here's Exactly How to Check If It's AI Generated

You got a message. Something feels off — too polished, arrived too fast, zero typos, no real personality. Here's how to check if it was AI generated, in under five minutes.

Why People Check Messages for AI

AI-generated messages show up everywhere now: customer service bots, dating app openers, phishing emails, or just someone using ChatGPT to reply faster. The reason you're checking matters less than knowing how to do it accurately. A single tool won't give you a verdict. Three steps together will.

Step 1: Scan It Manually for These Signs

Before running any tool, read the message once with these questions in mind. AI text has consistent tells:

• Perfect grammar, zero personality — humans leave micro-errors; AI usually doesn't
• Filler affirmations — phrases like "Absolutely!", "Certainly!", or "Great question!" appearing more than once
• Vague specifics — "I really enjoyed our conversation" with no reference to anything you actually said
• Even paragraph length — human messages are uneven; AI outputs are suspiciously balanced
• Overly complete sentences — every thought ties up neatly; real people leave loose ends

Two or three of these in one message is a strong signal. One alone means nothing.

Step 2: Run It Through a Detector

Paste the message into WriteMask's free AI detector. It analyzes perplexity and burstiness — the variance in sentence complexity that humans naturally produce and AI tends to flatten. No login needed. Results in seconds.

Short messages under 50 words are harder to score accurately. Anything over a paragraph gives you a reliable read. For context on what these scores actually mean, the explainer on how AI detectors work is worth two minutes of your time.

Step 3: Cross-Reference the Context

A detector score is evidence, not proof. Stack it against these questions:

• Did the reply arrive unusually fast?
• Does it address your specific points, or could it have been sent to anyone?
• Does the tone match earlier messages from this person?
• Is the sender in a role where they'd routinely use AI tools?

One high score across one message isn't a verdict. A pattern across multiple messages is.

What If the Detector Flags Something That's Clearly Human?

It happens more than people expect. Formal writers, non-native English speakers, and anyone who writes in structured or repetitive styles can score high on AI detectors. This is the AI detection false positive problem — documented, real, and worth understanding before you accuse anyone of anything.

If you're checking your own writing and it keeps coming back flagged, that's a separate situation. WriteMask has a 93% pass rate for humanizing AI text, but it also works the other direction — showing you what patterns in your writing are triggering detectors so you can adjust.

Quick Comparison: Tools for Checking Messages

• WriteMask Detect — free, no account, handles short-form content well
• GPTZero — stronger on longer documents, less reliable under 100 words
• Copyleaks — good for email-length content
• ZeroGPT — fast, but higher false positive rate on brief messages

For everyday message checks, the free AI detector is the fastest starting point with no friction.

The Short Answer

To check if a message is AI generated: scan manually for the signs above, run it through a detector, then weigh the score against context. None of these steps alone is reliable. All three together give you a confident read in under five minutes. Not sure how your own writing scores? The AI detection risk quiz shows you where you stand before anyone else runs a check on you.